Polymerized cyclodextrin microparticles for sustained antibiotic delivery in lung infections

Abstract

Pulmonary infections complicate chronic lung diseases requiring attention to both the pathophysiology and complexity associated with infection management. Patients with cystic fibrosis (CF) struggle with continuous bouts of pulmonary infections, contributing to lung destruction and eventual mortality. Additionally, CF patients struggle with airways that are highly viscous, with accumulated mucus creating optimal environments for bacteria colonization. The unique physiology and altered airway environment provide an ideal niche for bacteria to change their phenotype often becoming resistant to current treatments. Colonization with multiple pathogens at the same time further complicate treatment algorithms, requiring drug combinations that can challenge CF patient tolerance to treatment. The goal of this research initiative was to explore the utilization of a microparticle antibiotic delivery system, which could provide localized and sustained antibiotic dosing. The outcome of this work demonstrates the feasibility of providing efficient localized delivery of antibiotics to manage infection using both preclinical in vitro and in vivo CF infection models. The studies outlined in this manuscript demonstrate the proof-of-concept and unique capacity of polymerized cyclodextrin microparticles to provide site-directed management of pulmonary infections.

Keywords: Pseudomonas aeruginosa; cystic fibrosis; gentamicin; infection management; microparticle drug‐release; piperacillin; tazobactam; vancomycin.

Figure 1.. AutoDock Vina Validation.

Piperacillin and tazobactam both bind soluble cyclodextrin (CD) with increasing binding energies to $\alpha <\beta <\gamma$. Autodock Vina algorithm was used to simulate the binding of drug from CD to pCD. Dextran is chemically like pCD but lacks the affinity pocket, which is reflected in the lower binding energies, especially for the tazobactam. Each pCD type was done in triplicate to define the mean kCal/mol binding affinity for piperacillin and tazobactam.

Figure 2.. Piperacillin and Tazobactam Loading Efficiency.

Aqueous solutions of piperacillin and tazobactam (2 wt %, room temperature) were loaded into pCD for 72 h before extensive leaching in water and quantification via UV-Vis (Figure 2A). Piperacillin loading was significantly different by pCD type and crosslinking, which was driven by decreased loading at higher crosslinking. pCD-$\beta$ showed a small decrease between the two crosslinking densities tested (p = 0.052). pCD-$\alpha$ showed the largest (p = 0.013). Polymerized cyclodextrin $\gamma$ (pCD-$\gamma$) loads less piperacillin and tazobactam compared to pCD-$\alpha$ and pCD– $\beta$ (Figure 2B). Each pCD type was done in triplicate for two different experiments. Data is expressed in means±standard deviation.

Figure 2.. Piperacillin and Tazobactam Loading Efficiency.

Aqueous solutions of piperacillin and tazobactam (2 wt %, room temperature) were loaded into pCD for 72 h before extensive leaching in water and quantification via UV-Vis (Figure 2A). Piperacillin loading was significantly different by pCD type and crosslinking, which was driven by decreased loading at higher crosslinking. pCD-$\beta$ showed a small decrease between the two crosslinking densities tested (p = 0.052). pCD-$\alpha$ showed the largest (p = 0.013). Polymerized cyclodextrin $\gamma$ (pCD-$\gamma$) loads less piperacillin and tazobactam compared to pCD-$\alpha$ and pCD– $\beta$ (Figure 2B). Each pCD type was done in triplicate for two different experiments. Data is expressed in means±standard deviation.

Figure 3.. Piperacillin Release Kinetics.

Piperacillin released from pCD showed similar trends as drug loading. Piperacillin was loaded into pCD disks for drug loading (2 wt %, 72 h at room temperature) with release defined by using infinite sink conditions (1 ml per time point). Cumulative piperacillin release normalized to pCD weight (Figure 3A). Daily release of piperacillin from pCD. Piperacillin released from pCD at 1:0.16 was no longer detectable after 5 days (Figure 3B). Polymerized cyclodextrin at higher crosslinking (1:0.32) released piperacillin longer with a lower initial burst. Each crosslinking type was analyzed 2 times, with n=3 readings. Data is expressed in means±standard deviation.

Figure 3.. Piperacillin Release Kinetics.

Piperacillin released from pCD showed similar trends as drug loading. Piperacillin was loaded into pCD disks for drug loading (2 wt %, 72 h at room temperature) with release defined by using infinite sink conditions (1 ml per time point). Cumulative piperacillin release normalized to pCD weight (Figure 3A). Daily release of piperacillin from pCD. Piperacillin released from pCD at 1:0.16 was no longer detectable after 5 days (Figure 3B). Polymerized cyclodextrin at higher crosslinking (1:0.32) released piperacillin longer with a lower initial burst. Each crosslinking type was analyzed 2 times, with n=3 readings. Data is expressed in means±standard deviation.

Figure 4.. Tazobactam Release Kinetics.

Tazobactam released from pCD showed similar trends as drug loading. Tazobactam was loaded into pCD disks for drug loading (2 wt %, 72 h at room temperature) with a release run using infinite sink conditions (1 ml per time point). Cumulative tazobactam release normalized to pCD weight. pCD at higher crosslinking (1:0.32) released piperacillin for 12 days with a lower initial burst (Figure 4A). Daily release of tazobactam from pCD. Tazobactam released from pCD at 1:0.16 is only detectable for up to 24 hours (Figure 4B). Each crosslinking type was analyzed 2 times, with n=3 readings. Data is expressed in mean±standard deviation.

Figure 4.. Tazobactam Release Kinetics.

Tazobactam released from pCD showed similar trends as drug loading. Tazobactam was loaded into pCD disks for drug loading (2 wt %, 72 h at room temperature) with a release run using infinite sink conditions (1 ml per time point). Cumulative tazobactam release normalized to pCD weight. pCD at higher crosslinking (1:0.32) released piperacillin for 12 days with a lower initial burst (Figure 4A). Daily release of tazobactam from pCD. Tazobactam released from pCD at 1:0.16 is only detectable for up to 24 hours (Figure 4B). Each crosslinking type was analyzed 2 times, with n=3 readings. Data is expressed in mean±standard deviation.

Figure 5.. Piperacillin Microparticle Potency In Vitro.

Polymerized cyclodextrin type can be used to tailor bacterial killing kinetics. Piperacillin-loaded pCD cleared Gram-positive for between 5–10 days and Gram-negative bacteria for >21 days in an in vitro Zone of Inhibition assay. Drug-loaded disks were placed directly on freshly seeded exponential phase growth bacterial lawns and incubated for 24 h before measuring the zone with calipers. Polymerized cyclodextrin clearance of Staphylococcus aureus correlated with ring size and AutoDock Vina modeling: pCD-$\alpha$ cleared for 5 days pCD-$\beta$; for 8 days; and pCD-$\gamma$ for 10 days (Figure 5A). Piperacillin is utilized for Gram-negative coverage, which is reflected in both pCD-$\beta$ and pCD-$\gamma$ clearing Escherichia coli for >21 days (Figure 5B). In vitro, cyclodextrin potency analysis was done in triplicate against Staphylococcus aureus and Escherichia coli. Data is expressed in mean±standard deviation.

Figure 5.. Piperacillin Microparticle Potency In Vitro.

Polymerized cyclodextrin type can be used to tailor bacterial killing kinetics. Piperacillin-loaded pCD cleared Gram-positive for between 5–10 days and Gram-negative bacteria for >21 days in an in vitro Zone of Inhibition assay. Drug-loaded disks were placed directly on freshly seeded exponential phase growth bacterial lawns and incubated for 24 h before measuring the zone with calipers. Polymerized cyclodextrin clearance of Staphylococcus aureus correlated with ring size and AutoDock Vina modeling: pCD-$\alpha$ cleared for 5 days pCD-$\beta$; for 8 days; and pCD-$\gamma$ for 10 days (Figure 5A). Piperacillin is utilized for Gram-negative coverage, which is reflected in both pCD-$\beta$ and pCD-$\gamma$ clearing Escherichia coli for >21 days (Figure 5B). In vitro, cyclodextrin potency analysis was done in triplicate against Staphylococcus aureus and Escherichia coli. Data is expressed in mean±standard deviation.

Figure 6.. Gentamicin and Vancomycin Microparticle Potency In Vitro.

Gentamicin and vancomycin pCD-$\beta$ microparticles were evaluated against lawn of a clinical isolate of Pseudomonas aeruginosa. Both gentamicin (blue bars) and vancomycin (purple bars) microparticles had significant zonal effect against the pathogen with D2 and D3 of each antibiotic microparticles sets having the greatest impact (p<0.05). In vitro cyclodextrin potency analysis was done in triplicate against Pseudomonas aeruginosa. Data is expressed in mean±standard deviation.

Figure 7.. Lung Tissue Response to Microparticles.

Polymerized cyclodextrin microparticles were administered to WT mice to quantify mouse tolerance of pCD-$\beta$ microparticle size. Mice were weighed at baseline and then followed for 3 days for changes in clinical score and weight: gentamicin (purple), vancomycin (green) and empty pCD-$\beta$ microparticles (controls not containing antibiotics) (Figure 7A). Each mouse was its own control, data shows no impact on mouse weight over the course of treatment. Mice were euthanized for lung histology consistent with inflammation hematoxylin and eosin (H&E, Figure 7B 100x and Figure 7C 200x). The black arrows designate the airway epithelium and the interstitial tissue which demonstrates minimal inflammation. Periodic acid-Schiff was utilized to visualize mucin accumulatio (PAS, Figure 7D 100x and Figure 7E 200x). The black arrow designates the interstitial space which is clear with no deposition of excess mucins which would be pink fluid.

Figure 8.

Gentamicin containing microparticles were infused in the murine model of CF lung infection and inflammation using the same PAM5715 clinical isolate of Pseudomonas aeruginosa. The microparticles completely abolished pathogen load (A, p<0.05) while have a minimal effect on weight lost (B, p<0.05). Further, there was a trend towards decreased white blood cell recruitment into the lungs (C), suggesting an impact on aiding in inflammation resolution. Data is the result of n= 6 mice in each group, expressed as mean±standard deviation.